Assessment of Structural Brain Changes Related to Anoxic Coma Using High-field and Very Low Field Mobile MRI

Not Applicable

Not yet recruiting

• Conditions: Anoxic Coma; Cardiac Arrest (CA); Anoxia-Ischemia, Brain; Coma

• Registration Number: NCT07177755
• Lead Sponsor: University Hospital, Toulouse

Brief Summary

Standard predictors of outcome after cardiac arrest (CA) have substantial limitations in terms of reliability and generalizability. By providing brain structural connectivity maps, or connectomes, advanced MRI techniques, operating through high-strength magnetic field (HF; 1.5 to 3-T), have precisely revealed white and grey brain matter damages induced by CA, and have demonstrated the high sensitivity and specificity of these indicators for predicting neurological outcome after CA. However, HF MRI requires rigid safety precautions, highly trained technicians and patient transport to dedicated hospital imaging suites, hindering the implementation of these promising neuroimaging techniques in the setting of critical illness.

Interestingly, a recent report demonstrates the capability of a proof-of-concept very low-field (VLF; 0.064-T) mobile MRI to obtain neuroimaging at the bedside in critically ill patients. Nevertheless, the spatial resolution of VLF-MRI seems low and there is no available evidence about the use of VLF-MRI to extract highly needed new predictors of neurological recovery based on critical brain structural connectomes.

The CUBE project holds the promise of providing a radical paradigm shift in the field of neuroprognostication of anoxic coma patients. The current proposal is a "proof-of concept" study which aims to compare for the first time, HF, VLF and enhanced VLF (recon-VLF) structural connectomes from anoxic coma patients and healthy subjects across the time (3 paired HF and VLF brain scan across the first two weeks after CA). To obtain recon-VLF data, the Investigators will use an ensemble of ground-breaking methods to increase the native spatial resolution of VLF-MRI data. The whole brain imaging dataset will be used to prepare future neuroprognostication studies based on fully bedside assessment of brain structural integrity after CA.

Detailed Description

About two-thirds of patients who are in coma after resuscitation from cardiac arrest (CA) die before hospital discharge, of whom two-thirds die from neurological injury. However, only a minority of these deaths occur as a direct consequence of major brain injury. Actually, most of the deaths caused by CA-related hypoxic-ischemic brain injury result from withdrawal of life-sustaining treatments (WLST) following prognostication of a poor neurological outcome. Indeed, a timely, robust and generalizable neuroprognostication after CA appears to be essential to provide patient's family and caregivers with critical information to avoid both inappropriate WLST or prolonged treatment of patients with no chance of neurologically meaningful survival.

Many studies have investigated clinical examination, serum biomarkers, electrophysiological investigations, and neuroimaging test to predict neurological outcome after CA. Most of these tests have substantial limitation in terms of reliability, because many of them are based on low-certainty evidence due to their retrospective design, low power, potential bias (self-fulfilling prophecy) and confounders (sedation). For these reasons, to reduce prediction uncertainty, current guidelines recommend using a combination of predictors. Moreover, the optimal multimodal predictive strategy has yet to be defined and be prospectively validated and growing evidence suggest that these strategies can be applied only to a minority of patients.

In anoxic coma, prognostication is crucial in informing patient's relatives and to help clinicians to target treatments based on the patient's chances of achieving neurologically meaningful recovery. Currently, there is an urgent need for early, accurate and generalizable outcome predictors for these patients.

Although the exact brain neuronal structures subtending arousal and awareness are not yet perfectly delineated, some key subcortical and cortical regions have now been identified within an anterior forebrain 'meso-circuit' (comprising the reticular activating system, the striato-pallidal structures, the thalamus and associative cortical areas) and the 'frontoparietal' network. Accumulating evidence suggests that the common pathophysiological mechanism underlying coma is a broad withdrawal of excitatory synaptic activity within both these networks. Accordingly, the structural integrity of anterior forebrain 'mesocircuit' and the 'frontoparietal' networks have been consistently related to severity of chronic disorders of consciousness (vegetative state/unresponsive wakefulness syndrome - VS/UWS - and minimally conscious state/cortically mediated syndrome - MCS/CMS) and the potential of recovery from coma.

Overall, neurological recovery from coma is now conceptualized as being contingent upon the re-emergence of dynamic interaction between anterior forebrain 'mesocircuit' and 'frontoparietal' networks. Theoretically, assessing the structural integrity within these critical networks, hold the promise of providing innovative, powerful and mechanistic predictors of consciousness recovery after CA.

Several lines of evidence suggest that brain MRI can allow a timely and fine-grained evaluation of critical structural connectomes involved in consciousness recovery after CA. Indeed, brain MRI is currently recommended as the gold standard imaging tool for the evaluation of anoxic patients. For example, several studies have demonstrated that standard MRI is able to accurately assess brain intracerebral cytotoxic oedema after CA (hyperintensity on diffusion-weighted imaging - DWI - with corresponding low apparent diffusion coefficient - ADC - values). Patients with poor outcomes exhibited a nadir in ADC values at 3 - 5 days after CA, which therefore seemed to be the optimum time window for prognostication using ADC. However, there is no consensus regarding ADC thresholds, which extensively varied across a limited number of studies. Hence, standard MRI remains optional in most prognostication algorithm, probably because the low level of available evidence, an unmet need for methods standardization. These limitations have been a primary motivator for the deployment of advanced structural imaging techniques (sMRI), that can accurately provide highly needed information about the structural integrity of key cerebral connectomes after CA. Among the available methods, whole-brain white matter fractional anisotropy - WWM-FA -, measured by diffusor tensor imaging techniques - DTI and whole-brain gray matter morphometry - GMM- computed by voxel-based morphometry methods - VBM) seems to be the most promising.

Regarding the use of these MRI derived markers to reduce the incertitude of the neuroprognostication process after CA, the most compelling evidence come from two studies that have shown a high sensitivity and specificity of WWM-FA, and gray matter and GMM - measurements for predicting neurological recovery in anoxic patients. These two proof-of-concept studies suggest that such quantified evaluations of white and gray matter damage are able to predict poor neurological outcome more accurately than standard MRI. However, despite the promise hold by these advanced sMRI techniques, it is worth noting that these two studies collected data using poorly defined time window for MRI (2 - 6 weeks after CA) across lengthy data collection periods (\> 8 years) underlining the fact that due to inherent constraints of brain MRI, to date these methods can only be used as predictors of recovery in a small minority of anoxic comatose patients, are not able to provide predictive information when most needed, namely early after CA, and cannot be repeated over time.

Indeed, current guidelines and recommendations for the management of anoxic comatose patients recall the fact that HF MRI has practical limitations that hinder its use for clinically instable patients under life support assistance and continuous hemodynamic and respiratory monitoring. In fact, conventional MRI require strict, access-controlled environments, rigid safety precautions, highly trained technicians, and therefore entails patient transport to dedicated hospital imaging suites that have rendered MRI largely inaccessible for coma patients.

The Investigators have reported data that raises hope about the potential game-changing value of advanced structural brain MRI (sMRI) for neuroprognostication of comatose patients after CA. Hence, sMRI has demonstrated the usefulness of whole-brain WWM-FA and GMM to predict outcome after anoxic coma. Nevertheless, difficulties inherent to the use of conventional MRI in critically ill patients and ICU specific environments, have hindered the use of MRI to track brain anatomical changes across the time in anoxic patients and the implementation in clinical routine of these very promising neuroimaging techniques.

Conventional MRI (HF MRI) operates through high-strength magnetic fields (1.5 to 3 T) that require patients transfer to specific radiology facilities. In patients admitted to an intensive care unit (ICU), there are numerous risks involved in transportation to these imaging suites, including compromise of monitoring equipment, venous access limitations, and risk of endotracheal tube displacement. Moreover, critically ill patients carry often potential infectious pathogens that can impose considerable limitations on transportation to and decontamination of traditional imaging suites. Altogether these operational paradigms have rendered HF MRI largely inaccessible in the setting of critical care.

In this context, recent advances in very low-magnetic field MRI (VLF MRI), offer promise of safely enter MRI scanner in the critical care environment. A recent report demonstrates the capability of a proof-of-concept VLF (0.064-T) mobile MRI to obtain clinically meaningful brain imaging outside of radiology and in the presence of ferromagnetic material at the bedside of critically ill patients. It is worth noting that the availability of a bedside VLF MRI could constitute a radical paradigm shift in the field of neurological evaluation and early prognosis assessment of comatose patients. Indeed, VLF MRI appears to be a very promising alternative that might allow to obtain bedside neuroimaging for a larger number of critically ill patients (e.g., coma), at crucial timepoints that are still poorly explored (e.g., early after CA). Furthermore, the availability of a bedside MRI would also facilitate serial measurement, potentially providing unique insight about the time course of structural brain damage. These temporal profiles become particularly relevant for the design of potential interventional studies aimed to modulate brain plasticity and prevent neurological complications. The reported point-of-care VLF MRI used no cryogens and plugged into a single, 110-V, 15-A standard power outlet. The device dimensions rendered it manoeuvrable within the confines of an ICU patients' room. The 5-Gauss (0.0005-T) safety perimeter had a radius of 79 cm from the center of the magnet.

However, many gaps in the knowledge remains to be filled before implementing this new technology to coma neuroprognostication process.

The resolution of VLF MRI images acquired with routine clinical brain MRI sequences (T1-weighted, T1; T2-weighted, T2; T2-fluid-attenuated inversion recovery, FLAIR, diffusion-weighted, DWI and apparent diffusion coefficient, ADC) appears to be low and there is no available evidence about VLF MRI data reproducibility. Crucially, VLF MRI data have never been used to extract promising gray (WWM-FA) and white (GMM) matter predictive biomarkers for anoxic patients. Nevertheless, based on previous studies from the investigator group, the Investigators hypothesize that VLF MRI brain data carries critical information and might allow a radical paradigm shift in the field of neuroprognostication after CA.

The CUBE project aims to demonstrate that VLF MRIs can provide relevant patients' bedside information about both the initial impact of CA on brain anatomical architecture (i.e., primary brain injury) and the structural brain changes that may occur across the time (i.e., secondary brain injury). This information is crucial to prepare future prospective clinical studies that will specifically address the predictive value VLF MRI brain data after CA and will allow the design of innovative personalized therapeutics for anoxic comatose patients.

Study Timeline

• Study First Submitted: 2025-05-20
• Status Verified: 2025-09
• Study First Submitted QC: 2025-09-10
• Last Update Submitted: 2025-09-10
• Study Start: 2025-09-15
• Study First Posted: 2025-09-17
• Last Update Posted: 2025-09-17
• Primary Completion: 2028-09
• Study Completion: 2029-03-01

Recruitment & Eligibility

• Status: NOT_YET_RECRUITING
• Sex: All
• Target Recruitment: 60

Inclusion Criteria

Not provided

Exclusion Criteria

Not provided

Study & Design

• Study Type: INTERVENTIONAL
• Study Design: PARALLEL

Primary Outcome Measures

Name	Time	Method
Differences between whole-brain white (WWM-FA) and grey (GMM) matter global metrics independently acquired by HF and VLF-MRI scans very early after coma onset (D0)	Day 0	White matter will be assessed by computing the White Matter Fractional Anisotropy (WWM-FA), which ranges from 0 to 1 with lower values related to worse outcomes. Grey matter will be evaluated using Global brain Matter Morphometry (GMM), which encompass both cortical thickness (mm) and deep grey matter volumetry (mm3) assessments, with lower values being associated with poorer outcomes. Both WWM-FA and GMM data will be normalized using brain MRI data obtained from healthy controls (z-scores).

Secondary Outcome Measures

Name	Time	Method
Differences between HF and VLF white (WWM-FA=White Matter Fractional Anisotropy ) and grey (GMM=Global brain Matter Morphometry) matter z-scores.	Day 0	White matter will be assessed by computing the White Matter Fractional Anisotropy (WWM-FA), which ranges from 0 to 1 with lower values related to worse outcomes.; Grey matter will be evaluated using Global brain Matter Morphometry (GMM), which encompass both cortical thickness (mm) and deep grey matter volumetry (mm3) assessments, with lower values being associated with poorer outcomes.; Both WWM-FA and GMM data will be normalized using brain MRI data obtained from healthy controls (z-scores).
Differences between HF and VLF white (WWM-FA=White Matter Fractional Anisotropy ) and grey (GMM=Global brain Matter Morphometry) matter z-score	Day 15	White matter will be assessed by computing the White Matter Fractional Anisotropy (WWM-FA), which ranges from 0 to 1 with lower values related to worse outcomes.; Grey matter will be evaluated using Global brain Matter Morphometry (GMM), which encompass both cortical thickness (mm) and deep grey matter volumetry (mm3) assessments, with lower values being associated with poorer outcomes.; Both WWM-FA and GMM data will be normalized using brain MRI data obtained from healthy controls (z-scores).
Paired comparison between recon-VLF, VLF and HF MRI image's quality, using radiometric information metrics (PSNR, SSIM, ERGAS)	day 15	PSNR (Peak Signal-to-Noise Ratio) is a ratio between the maximum possible signal and the noise. The PSNR theoretically can approach infinity if there is no difference between the compared images and go down to 0, meaning in that case that there is significant distortion between the visual dataset. The SSIM (Structure Similarity Index) considers changes in structural information, luminance, and texture, making it more aligned with human visual perception. The maximum value of SSIM is 1, which indicates that the two images are identical. The minimum SSIM value is 0, which means no similarity between the two images. The ERGAS (Error-Related Global Accuracy Statistic) evaluates the quality of super-resolution tasks. The minimum value for ERGAS is 0, which indicates perfect fusion with the reference image. The maximum value for ERGAS is theoretically infinity, as the error can be arbitrarily large if there is a significant mismatch between the compared images.
correlation between HF, VFL and recon-VLF normalized WWM-FA and GMM metrics (z-scores) collected across the first two weeks after ROSC and patient's neurological performance at 6 months according to the modified Rankin Score (mRS, ranking from 0-6).	month 6	Association between HF, VFL and recon-VLF normalized WWM-FA and GMM metrics (z-scores) collected across the first two weeks after ROSC and patient's neurological performance at 6 months according to the modified Rankin Score (mRS, ranking from 0-6).; The mRS which will be dichotomized as follow: 0-3 (favourable outcome) and 4-6 (unfavourable outcome).
correlation between HF, VFL and recon-VLF normalized WWM-FA and GMM metrics (z-scores) and patient's consciousness level at 6 months according to the Consciousness Recovery Scale-Revised (CRS-R, ranking from 0 to 23).	month 6	Association between HF, VFL and recon-VLF normalized WWM-FA and GMM metrics (z-scores) and patient's consciousness level at 6 months according to the Consciousness Recovery Scale-Revised (CRS-R, ranking from 0 to 23). The CRS-R evaluates different points assigned based on how well the patient responds to specific stimuli. Despite the absence of threshold, a higher score indicates a higher level of awareness, while a lower score suggests more significant impairment quality of life
correlation between HF, VFL and recon-VLF normalized WWM-FA and GMM metrics (z-scores) and patient's quality of life at 6 months according to the Health-Related Quality of Life after Cardiac Arrest (QOL-CA, ranking from 0 to 100).	month 6	Association between HF, VFL and recon-VLF normalized WWM-FA and GMM metrics (z-scores) and patient's quality of life at 6 months according to the Health-Related Quality of Life after Cardiac Arrest (QOL-CA, ranking from 0 to 100). The QOL-CA is a measure of a person's well-being, encompassing physical, mental, and societal health. A QOL-CA score of 100 represents the best possible quality of life, while a score of 0 would indicate the worst possible outcome.

Trial Locations

Locations (1)

CHU de Toulouse; 🇫🇷 Toulouse, France